<p>Social media platforms like X (Twitter) have become indispensable tools for health promotion, particularly during global crises like the COVID-19 pandemic, by facilitating vital communication among the public, researchers, and health professionals. These platforms enable the viral spread of information, thereby amplifying its impact and potentially improving health outcomes. Misinformation can also spread rapidly, complicating efforts to ensure the public receives reliable health guidance. Therefore, understanding how health-related tweets propagate is crucial for addressing this challenge and counteracting misinformation. This study introduces an innovative Inverse Reinforcement Learning (IRL) approach to analyze factors influencing the popularity of health-related tweets, using the #onehealth hashtag as a case study. We designed a learning framework that emulates the X ecosystem, developed an IRL model to identify factors influencing retweets, and computed an optimal policy for maximizing retweeting behavior. By using expert tweets as a benchmark and their number of retweets as a reward, our results revealed that while certain content elements (e.g., URLs, international day mentions, or hashtags) and timing are important factors influencing tweet popularity, tweet impact can be nuanced and context-dependent. Our model provided a strategic roadmap for users to maximize their tweet impact, guiding them on when to intensify or decrease efforts to maintain optimal engagement with health-related information. The IRL approach introduced here can be adapted and applied to diverse social media contexts and a wider range of topics, reinforcing the versatility and significance of our method in enhancing social media strategy and engagement across various platforms and thematic areas.</p>